Embed Size (px)
Citation preview
Generalized Taylor and Generalized Calvo Price and Wage-Setting: Micro Evidence
with Macro Implications
Huw Dixon Hervé Le Bihan
CESIFO WORKING PAPER NO. 3119 CATEGORY 7: MONETARY POLICY AND INTERNATIONAL FINANCE
JULY 2010
An electronic version of the paper may be downloaded • from the SSRN website: www.SSRN.com • from the RePEc website: www.RePEc.org
• from the CESifo website: Twww.CESifo-group.org/wp T
CESifo Working Paper No. 3119
Generalized Taylor and Generalized Calvo Price and Wage-Setting: Micro Evidence
with Macro Implications
Abstract The Generalized Calvo and the Generalized Taylor model of price and wage-setting are, unlike the standard Calvo and Taylor counter-parts, exactly consistent with the distribution of durations observed in the data. Using price and wage micro-data from a major euro-area economy (France), we develop calibrated versions of these models. We assess the consequences for monetary policy transmission by embedding these calibrated models in a standard DSGE model. The Generalized Taylor model is found to help rationalizing the hump-shaped response of inflation, without resorting to the counterfactual assumption of systematic wage and price indexation.
JEL-Code: E31, E32, E52, J30.
Keywords: contract length, steady state, hazard rate, Calvo, Taylor, wage-setting, price-setting.
Huw Dixon Cardiff Business School
Colum Drive UK – Cardiff CF10 3EU
United Kingdom [email protected]
Hervé Le Bihan Banque de France
Direction des Etudes Microéconomiques et Structurelles / SAMIC
31 rue Croix des Petits Champs F – 75001 Paris
France [email protected]
June 25, 2010 The authors thank Julien Matheron for helpful remarks. They are grateful to Michel Juillard for help in simulating our model with the Dynare code. The views expressed in this paper may not necessarily be those of the Banque de France. Huw Dixon thanks the Fondation Banque de France for funding his participation in this research.
1 Introduction
Christiano, Eichenbaum and Evans (2004) (hereafter CEE) and Smets andWouters (2003) (SW ) have developed Dynamic Stochastic General Equilib-rium models of the US and euro area economies that have become standardtools for monetary policy analysis. These models have been designed to re-�ect the empirical properties of the US and euro area data in a way that isconsistent with New Keynesian theory. In particular these models have beenshown to replicate the impulse-response functions of output and in�ation toa monetary policy shock. Central to these models is the Calvo model of priceand wage setting with indexation developed by Erceg, Henderson and Levin(2000)(EHL): �rms (unions) have a constant probability to be able to opti-mally reset prices (wages); when �rms (unions) do not optimally reset prices(wages), the nominal price (wage) is automaticaly updated in response toin�ation.1 This approach is however inconsistent with the micro-data alongtwo dimensions. First, it assumes that the probability of price reoptimizationis constant over time. Second, it implies that nominal wages and prices adjustevery period, which is counterfactual as noted e.g. by Cogley and Sbordone(2008) and Dixon and Kara (2010).The purpose of this paper is to take seriously the recent micro-data ev-
idence on wages and prices and apply it directly to alternative wage andpricing models. Our main point of departure is the aggregate distribution ofdurations of price and wage spells. In steady-state, this can be representedin three di¤erent ways: the Hazard pro�le, the distribution of durations,and the cross-sectional distribution (see Dixon 2009 for a detailed explana-tion). We take the Hazard pro�le and use this to calibrate a GeneralizedCalvo (GC) model with duration-dependent reset probabilities.2 We takethe cross-sectional distribution of completed spells and use this to calibratea Generalized Taylor Economy (GTE) in which there are several sectors,each with a simple Taylor contract but with contract lengths di¤ering acrosssectors3. Each of the two models we consider (GC and GTE) exactly re�ectsthe full distribution of durations revealed by the micro-data. We also consider
1In EHL, the indexation is to the unconditional mean in�ation, while in SW and CEEit is to lagged in�ation
2The GC approach has been adopted by Wolman (1999), Guerrieri (2006), Dixon(2009).
3References for the GT price setting model include Taylor (1993), Dixon and Kara(2005, 2010), Coenen et al. (2007).
2
the simple Calvo model with the reset probability calibrated by the averageproportion of wages or prices changing in the data.In order to carry out a quantitative experiment, we use original micro data
on wages and prices in France. Whilst the data on prices has been well studiedfor a range of countries (Dhyne et al. 2006, Klenow and Malin, 2010), relevantwage data are harder to �nd. We are here able to use a unique, quarterlydata set on wages from France (Heckel, Le Bihan, Montornes, 2008). Ourapproach is then to substitute the standard Calvo scheme with one basedon the micro-data using the GC and GTE pricing models and investigatehow far these approaches work when set in the SW model of the euro areaeconomy. While we use data for one country of the euro area (France), wewould argue they are a relevant proxy for the whole euro area, for whichsimilar hazard function are not available. Comparative evidence for pricesdoes indeed suggest that there is a large degree of similarity across the largereuro area economies (Dhyne et al. 2006). Finally, we are able to study macrodynamics, in particular the response to a monetary policy shock.With respect to previous research that has used GC or GT models (e.g.
Wolman 1999, Coenen et al. 2007, Dixon and Kara 2010, Kara 2010), ourspeci�c contribution is twofold. First, we use direct evidence on the actualdistribution of both wages and price durations. By contrast, previous researchhas used either only a few moments of these distributions or indirectly es-timated distributions. Second we derive a model of wage-setting with GTand GC contracts, which builds on the EHL model. This extends the EHLframework to a more general and �exible structure of wage-rigidity than hasbeen considered previously.Our exercise is to a large extent an analytical one: the SW and CEE
models and their clones rely on indexation to generate some of the featuresthat make the models congruent with the macro-data: in particular, thedegree of persistence in output and in�ation in response to monetary shocksand the "hump shape" found in the macro-data. Since indexation is largelyat odds with the micro-data, we want to see how far we can go keepingthe SW=CEE framework but replacing indexation with a more rigorouslymicro-data based approach to pricing. Our main result is that using thesealternative frameworks we can partly replicate the persistence of in�ationand output following shocks without relying on indexation. In particular theGeneralized Taylor model is shown to be able to produce a hump-shapedresponse of in�ation and output to monetary policy shocks, which does nothappen with the Calvo based approaches. In contrast, we �nd that all three
3
approaches lead to similar responses to a productivity shock.The structure of the paper is as follows. Section 2 develops GT and
GC models of price and wage setting. Section 3 presents our micro dataon price and wages and uses the distribution of durations to calibrate thesemodels. Section 4 embeds these calibrated GC and GT price and wage-setting schemes into the Smets and Wouters model of the euro area economy,and studies the implications for the monetary policy transmission mechanism.Section 5 concludes.
2 Price and Wage -setting in GT and GCeconomies
Standard time-dependent models of price rigidity have restrictive implica-tions for the distribution of durations. The standard Taylor model predictsthat all durations are identical. The standard Calvo (constant hazard) modelpredicts that durations are distributed according to the exponential distri-bution. In this paper, we consider the Generalized Taylor and GeneralizedCalvo set-ups which allow the distribution of durations implied by the pric-ing model to be exactly the same as the distribution found in the actualmicro-data. The distribution of durations can be characterized in variousways. As shown in Dixon (2009), in steady-state there are a set of iden-tities that link the Hazard function and the cross-sectional distribution ofcompleted contracts lengths. These are just di¤erent ways of looking at thesame data. However, the Hazard function relates naturally to the GeneralizedCalvo model where the hazard rates are mapped on to duration dependentprice-reset probabilities. The cross-section of completed price-spell lengthsis easily related to the Generalized Taylor model, where there are many sec-tors, and within each sector there is a simple Taylor staggered contract whichdi¤er across sectors.We will �rst outline the Generalized Taylor and Generalized Calvo economies
in terms of price-setting behavior. We will then see how this applies to wage-setting.
2.1 Generalized Taylor Economy (GTE)
In the Generalized Taylor Economy (GTE) there are N sectors, i = 1; :::; N:In sector i there are i�period contracts: each period a cohort of i�1 of the
4
�rms in the sector sets a new price (or wage). If we think of the economyas a continuum of �rms, we can describe the GTE as a vector of sectorshares: �i is the proportion of �rms that have price-spells of length i. Ifthe longest observed price-spell is F , then we have
PFi=1 �i = 1 and � =
(�1; :::; �F ) is the F -vector of shares. We can think of the "sectors" as"duration sectors", de�ned by the length of price-spells. The essence of theTaylor model is that when they set the price, the �rm knows exactly howlong its price is going to last. The simple Taylor economy is a special casewhere there is only one length of price-spell (e.g. �2 = 1 is a simple Taylor"2 quarter" economy). The GTE is based on the cross-sectional distributionof completed spell lengths: hence it can also be called the distribution across�rms (DAF ) in this context. The GTE has been developed in Taylor (1993),Carvalho (1995), Dixon and Kara (2005, 2006, 2010), Coenen et al (2007)and Kara (2010). The GTE can represent any steady-state distribution ofdurations: hence it can be chosen to exactly re�ect the distribution found inthe micro-data.The log-linearised equation for the aggregate price pt is a weighted average
of the sectoral prices pit, where the weights are �i :
pt =FXi=1
�ipit (1)
In each sector i, a proportion i�1 of the �i �rms reset their price at eachdate. Assuming imperfect competition and standard demand curve, the op-timal reset price in sector i; xit is given by the �rst-order condition of anintertemporal pro�t-maximisation program under the constraint implied byprice rigidity. The log-linearised equation for the reset price, as in the stan-dard Taylor set-up, is then given by :
xit =
1Pi�1k=0 �
k
!i�1Xk=0
�kEtp�t+k (2)
where � is a discount factor, Et is the expectation operator conditional oninformation available at date t , and p�t+k is the optimal �ex price at timet+ k. The reset price is thus an average over the optimal �ex prices for theduration of the contract (or price-spell). The formula for the optimal �exprice will depend on the model: clearly, it is a markup on marginal cost. Wewill specify the exact log-linearised equation for the optimal �ex-price whenwe specify the precise macreconomic model we use.
5
The sectoral price is simply the average over the i cohorts in the sector:
pit =1
i
i�1Xk=0
xit�k (3)
In each period, a proportion �h of �rms reset their prices in this economy:proportion ��1 of sector i which is of size �i:
�h =
FXi=1
�ii
(4)
2.2 The Generalized Calvo Economy (GCE)
In the Generalized Calvo Economy (GCE), initially developed by Wolman(1999), �rms have a common set of duration-dependent reset probabilities:the probability of resetting price i periods after you last reset the price is givenby hi. This is a time-dependent model, and the pro�le of reset probabilitiesis h = fhigFi=1. Clearly, if F is the longest price-spell we have hF = 1 andhi 2 [0; 1) for i = 1:::F � 1. Again, the duration data can be represented bythe hazard function. Estimated hazard function can then be used to calibrateh. Since any distribution of durations can be represented by the appropriatehazard function, we can choose the GCE to exactly �t micro-data.In economic terms, the di¤erence between the Calvo approach and the
Taylor approach is that when the �rm sets its price, it does not know howlong its price is going to last. Rather, it has a survivor function S(i) whichgives the probability that its price will last at up to i periods. The survivorfunction in discrete time is4:
S(1) = 1 (5)
S(i) =i�1Yj=1
(1� hj) i = 2; :::; F
Thus, when they set the price in period t, the �rms know that they will lastone period with certainty, at least 2 periods with probability S(2) and so
4Note that the discrete time survivor function e¤ectively assumes that all "failures"occur at the end of the period (or the start of the next period): this corresponds tothe pricing models where the price is set for a whole period and can only change at thetransition from one period to the next.
6
on. The Calvo model is a special case where the hazard is constant hi = �h,S(i) = (1��h)i�1 and F =1. Of course, in any actual data set, F is �nite. Inthe applications which follow we set F = 20 quarters, close to the maximumduration observed in price micro data.In the GC model the reset price is common across all �rms that reset
their price. The optimal reset price, in the same monopolistic competitionset-up as mentioned above, is given in log-linearised form by:
xt =1PF
i=1 S(i)�i�1
FXi=1
S(i)�i�1Etp�t+i�1 (6)
The evolution of the aggregate price-level is given by:
pt =FXi=1
S(i)xt�i+1 (7)
That is, the current price level is constituted by the surviving reset prices ofthe present and last F periods.
2.3 Wage-setting.
We can apply GCE and GCT to wage data in order to calibrate wage-setting.If we have a model with �exible prices, simply using the same equations asthe price-setting model would probably be a relevant shortcut. Indeed aswas shown in Ascari (2003) and Edge (2002), models of either wage or pricerigidity lead to reduced-form dynamics that is largely similar for reasonableparameter values. So, calibrating the models of sections 2:1 and 2:2 withthe distributions implied by the wage data would presumably be a relevantstrategy.However, we also wish to provide a model that combines both wage and
price rigidity as in the models of Erceg et al. (2000), Christiano et al. (2005),Smets and Wouters (2003). Clearly, the description of pricing decisions de-scribed above will continue to hold. What we need to add are the speci�cequations for marginal cost with sticky wages. As in EHL, we take the craft-union model �rst employed in the macroeconomic setting by Blanchard andKiyotaki (1987). In this case, there is a CES aggregator for labour inputswith a speci�c elasticity �w. There is a unit interval of households h 2 [0; 1]
7
each with a unique type of labour. Aggregate labour Lt is constituted of bycombining each household�s labour Lt(h) according to:
Lt =
�Z 1
0
Lt(h)�w�1�w dh
� �w�w�1
The corresponding aggregate unit wage-cost index is derived from individualhousehold wages Wt(h)
Wt =
�ZWt(h)
1��wdh
� 11��w
where �w is the elasticity of the corresponding conditional labour demand:
Lt(h) =
�Wt(h)
Wt
���wLt (8)
We assume that the household preferences are described by the followingutility function that features habit formation
E0
1Xt=0
�tU(Ct �Ht; 1� Lt(h))
where Ht = bCt�1; b is a parameter describing habit formation, assumedto be external, and Lt(h) is hours worked by household h. We specify thefunctional form for U as:
U(Ct �Ht; 1� L(h)t) =1
1��c(Ct �Ht)
1��c +1
1� �L(1� Lt(h))
1��L
where �c is the inverse of intertemporal elasticity of substitution, and �L isthe inverse of the elasticity of hours worked to the real wage rate.We assume full-insurance so that the level of consumption will be equal
across households5. Employment is assumed to be demand determined:hence the households marginal rate of substitution at time t is:
MRS(h)t = �Ul(Ct � bCt�1; 1� L(h)t)
UC(Ct � bCt�1; 1� L(h)t)=(Ct � bCt�1)
�c
(1� Lt(h))�L(9)
5See Ascari (2000) for the details.
8
The union-household sets its nominal wage W (h)t. We can de�ne the"shadow nominal wage" as:
W �(h)t = Pt:MRS(h)t (10)
W �(h)t is nominal wage which would equate the real wage with the marginalrate of substitution for household h given the labour which is demanded ofit at its current nominal wage W (h)t (from 8), and its current and pastconsumption according to (9).
2.3.1 Wage-setting GTE.
Log-linearising these equations (9),(8),(10) we have:
mrs(h)t = �Ln(h)t +�c1� b
(ct � b:ct�1) (11)
n(h)t = �w (wt � w(h)t) + nt (12)
w�(h) = pt +mrst (13)
where lowercase letter are log-deviation and n(h)t is the log-deviation ofLt(h): If the household-union knows the length of its contract to be i periods,the (nominal) reset wage xwit will ful�ll w(h)t+k = xit for k = 0; :::; i � 1.The optimal reset wage is obtained by maximizing the intertemporal utilityfunction subject to this structure of wage stickiness, and a standard budgetconstraint. In log-linear form the optimal reset wage is given by:
xwit =
1Pi�1k=0 �
k
!i�1Xk=0
�kEtw�t+k (14)
That is, xwit is a weighted average of the discounted nominal shadow wagesw�t+k.As shown in the appendix, using equations (11),(12),(13) it is straight-
forward to derive the reset wage equation:
xwit =1
(1 + �L�w)Pi�1
k=0 �k
i�1Xk=0
�kEt
�pt+k + �L (�wwt+k + nt+k) +
�c1� h
(ct+k � b:ct+k�1)
�(15)
9
Therefore we can construct a wage setting GTE: The aggregate wage isrelated to the sectoral wages wit; where the weights �iw come from the cross-sectional distribution across �rms in the data. The sectoral wages wit aresimply an average across past reset wages in that sector:
wt =
FwXi=1
�iwwit (16)
wit =1
i
i�1Xk=0
xwit�k (17)
These equations can then be combined with the price-setting GTE equationsto simulate an economy with GT nominal rigidity in both price and wagesetting. Clearly, the wage-setting decision will depend directly on the levelof the aggregate variables (Lt; Ct) and indirectly on the rest of the variablesin the model.
2.3.2 Wage-setting GCE.
In the case of the GCE, we have the wage-survival function and related hazardrates: Sw(i) and hw(i) i = 1; :::; Fw derived from the data on wages. Theoptimal reset wage is the same for all �rms, and is given by the log-linearized�rst order condition:
xwt =1PF
i=1 Sw(i)�i�1
FwXi=1
Sw(i)�i�1Etw
�t+i�1 (18)
=1
(1 + �L�w)PFw
i=1 Sw(i)�i�1
FwXi=1
�i�1Et(pt+i�1 (19)
+�L (�wwt+i�1 + nt+i�1) +�c1� h
(ct+i�1 � b:ct+i�2))
The aggregate wage is an average of past reset prices, weighted by survivalprobabilities:
wt =FwXi=1
Sw(i)xt�i+1 (20)
Again, this wage-setting GCE can be combined with price-rigidity. Notethat we can treat the Calvo model as a special case of the GCE. We can
10
use the average proportion of wages reset each quarter as our calibration ofthe Calvo reset probability: the resulting GCE is a constant hazard modelhw(i) = �hw for i = 1:::Fw. In practice, we truncate the wage setting to amaximum duration of 20 quarters, rather than having the in�nite horizonassumed by the theoretical Calvo model. The truncation at Fw = 20 hasalmost no quantitative impact on the conclusions derived from the modelgiven that in our data �hw = 0:38. Removing the in�nite time horizon mayin any case be seen an improvement on the Calvo model.Note that in the case of the constant hazard, equation, combining (19)
and (20) yields the "new Keynesian Phillips curve" formulation found inSW 6, which writes the wage-setting equation in terms of price in�ation, wagein�ation and the sum of current and future deviations of the real wage fromtheMRS between consumption and leisure. Equation (19) is probably moreintuitive and easy to understand than the NKPC-like formulation. Note alsowe have log-linearized the model around a zero in�ation rate steady-state (asis the case in the NKPC formulations of CEE and SW ) which means that thewage and price levels are stationary: if there was non-zero in�ation in steady-state, this would not be the case. However, as Ascari (2004) demonstrates,this also invalidates traditional formulations of the NKPC.
3 The hazard function of price and wage changes:micro evidence
This section describes the micro data we use to characterize the distributionof wages and prices, and report some important statistics about this distri-bution. We con�ne ourselves to a brief description, since a more completedescription and details can be found in earlier papers.
3.1 Data
The dataset used in the case of prices is composed of the consumer pricequotes collected by the INSEE, the French Statistical Institute, to buildthe CPI (Consumer Price Index). A detailed investigation of this dataset ispresented in Baudry et al. (2007). The sample contains around 13 millionprice observations collected monthly over the 9 year period 1994:7 to 2003:2.
6See SW equation (33) page 1138.
11
Data are available for a range of goods that cover 65% of the French CPIdata. These data are collected for several hundreds of elementary products,at di¤erent outlets and at di¤erent months. An individual observation is aprice quote Pjkt for product j at outlet k at time t (t=1....104). The resultingdataset is a panel with about 125,000 price quotes each of the 104 months.The panel is unbalanced since the range of products and the outlets arechanged over time for reasons to do with constructing the CPI. The datasetalso includes CPI weigths, which we use to compute aggregate statistics.From the panel of prices, we can compute the frequency of price changes,i.e. the average proportion of prices that do change a given month. Onour sample this weighted average frequency is equal to 19%: this statistic isthe empirical counterpart of the Calvo parameter in discrete time. This is amonthly statistic: it corresponds to the quarterly frequency of �h = 0:53:Consistently with the concepts introduced in section 2, we can organize
this data into price spells. These are a sequence of price-quotes at the sameoutlet for which the price quoted is the same. There are 2,372,000 pricespells in the panel The weighted average duration of price spells is 7.2months.7 There are several data issues, which are discussed in Baudry etal. (2007). Not least is the issue of censored data: we can have left truncateddata, where the beginning of the price spell is not observed. We have righttruncated data, where we do not observe the end of the spell. We alsoobserve spells which are both right and left truncated: we know neither thebeginning or the end. Truncation results either from the turn-over of productsin stores, and from changes in the sample decided by the statistical institute.The majority of price spells are uncensored: 57%. There are a lot of lefttruncated spells: 27%. The rest are either right truncated or truncated atboth ends. In our empirical analysis below we will focus on the distributionof spells that are non-left-censored (and disregard other spells). We includeright-truncated spells (i.e. price trajectories that are terminated before theactual end of sample) because we interpret them as completed spells: forexample we regard product substitution in a store as actually ending a pricespell. There are of course di¤erent ways of interpreting truncation. However,we have carried out our analysis using alternative treatments of censoringand our results were robust.
7The maximum duration in the dataset is 104 months, but this concern a negligiblefraction of price spells. The model simulations that follow use a truncation of the hazardfunction at F = 20 quarters. This has no material empirical consequence since less than0.03 percent of price spells last more than 60 months.
12
To characterize the distribution of wage durations, we here rely on asurvey of �rms conducted by the French Ministry of Labour, the ACEMOsurvey. The ACEMO is unique, owing to its quarterly frequency. Indeed,while CPI data are collected at the monthly frequency in a very standard-ized fashion for many countries, data on wages at a higher frequency thanannual are scarce. The ACEMO dataset is analyzed in Heckel, Le Bihanand Montornes (2008). The ACEMO survey covers establishments with atleast ten employees in the non-farm market sector. Data are collected at theend of every quarter from a sample of about 38,000 establishments. Theavailable �les span the period from the fourth quarter of 1998 to the fourthquarter of 2005. The ACEMO survey collects the level of the monthly basewage, inclusive of employee social security contributions. The data excludesbonuses, allowances, and other forms of compensations. The survey collectsthe wage level of representative employees, for four categories of positionswithin the �rm: manual workers, clerical workers, intermediate occupations,managers. Each �rm has to report the wages level of up to 12 employees,representative of the four above mentioned occupations (1 to 3 occupationsin each category). Measurement error is a crucial concern when analyzingwage data. Here, this concern is attenuated because we have answers by�rm to a compulsory survey, rather than self-reported household answers asin many studies. Furthermore the statistical agency performs some qualitychecks. The data set contains some information which allows us to make surethat the individuals are actually the same from one quarter to another.The �nal dataset contains around 3.7 million wage records and around
1.8 million wage spells. To produce aggregate statistics, data are weightedusing the weight of �rms and sectors in overall employment. The averagefrequency of wage change is 38% per quarter (�hw = 0:38), while the weightedaverage duration of spells is 2.0 quarters. Less than 0.1 percent of wage spelllast more than 16 quarters.8
3.2 Hazard function estimates
From the weighted distribution of price and wage durations, we compute sur-vival function and hazard functions using the non-parametric Kaplan-Meierestimator. The estimates of the hazard function, the parameters hi of sec-
8In the model simulations we use a truncation of the hazard function at a maximumduration of Fw = 20 quarters. Virtually no information is thus lost.
13
tion 2.2, are presented in Figure 1.9 Importantly, note that the hazard func-tion for prices relates to monthly data while that for wages relates quarterlydata, consistent with the original frequency of the data. When proceeding tomodel-based analysis below, information on price spells will be converted tothe quarterly frequency. As discussed above, these hazard functions whereobtained by discarding left-censored spell and treating right-censored spellsas a price or wage changes, but our results are robust to other assumptionson censoring.
Insert FIGURE 1
The hazard function for prices is typical of that observed in recent researchwith micro price data (see Dhyne et al., 2006, Klenow and Malin, 2010). Ittends to be decreasing over the �rst months. This, to some extent, re�ectsheterogeneity across sectors in the baseline level of price rigidity (see Alvarezet al., 2005, Fougère et al, 2007 for a discussion and empirical investigations).There is a massive spike at duration 12 months, indicating that a lot ofretailers change their prices after exactly 1 year. The hazard function of wageis �atter than prices, but clear spikes are seen at duration 4 and 8 quarters.Overall, the bottomline for both price and wage is that hazard functions areneither �at (as the simple Calvo model would predict), nor degenerate spikesat a given duration (as in the Taylor model), but have a more general shapethat mixes patterns of these two cases. We view these observed patterns as amotivation for using Generalized Taylor and Generalized Contracts to re�ectthe estimated distributions.The two panels of Figure 2 present the distribution of durations, as well
as the Distribution across Firms (i.e. the parameters �i and �iw de�nedin sections 2.1. and 2.3.1), for prices and wages respectively. These �guresconvey the same information as the hazard function. They make more vis-ible that at at given date, the cross-section of spells is dominated by �rmsthat experience a one-year price or wage contract. For wages, one observesthat there is a substantial mass of short durations, which explain why theaverage duration for wages is rather short. This observation does not com-pletely conform with intuition and requires some quali�cations. Following
9Due to the huge number of observations, con�dence intervals are very narrow, thusare not reported. The �gure contains the estimates for the �rst 16 months, although weestimated the hazard function for F = 95 (IS this correctXXX). Details available fromthe authors.
14
Heckel et al (2008), our interpretation is that this result re�ects to a largeextent cases where one single decision of wage increase (say a yearly generalincrease in a given �rm) is spread out over the year and split up between twoor (more) smaller wage increases10. Informal evidence suggest that a fractionof French �rms actually follow such a policy of gradual implementation ofwage increase. The prevalence of such a pattern is con�rmed by the empir-ical analysis of wage-agreement data by Avouyi-Dovi, Fougère and Gautier(2010). For a given duration of wages, these types of cases create moreinertia than the one predicted by sticky wage models, because some wagechanges are based on past information (as in Mankiw and Reis, 2002). Theyare thus pre-determined and cannot respond to current shocks. While it isdi¢ cult to correct for the degree of such pre-determination in our dataset,we simply note that our duration measures, and thus our model-based analy-sis, may tend to underestimate the degree of wage rigidity, and presumablymacroeconomic persistence.
Insert FIGURE 2
4 Implications for monetary policy transmis-sion.
In this section, we use the distribution of the price and wage data to calibratetheGT andGC models developed in section 2. We then embed these model intwo alternative macroeconomic models to investigate the implications of GCand GT behavior for in�ation and output persistence following a monetarypolicy shock.
4.1 A simple quantity theory model with price or wage-setting.
We will �rst examine the GC and GTE models of prices in a quantity theorymodel with labour as the only input of production. This model has the greatadvantage of being very simple, because almost all its dynamic properties aregenerated by the pricing models alone. DSGE models like the SW model in
10In e¤ect, this behaviour is similar to the Fischer-like contracts used in sticky-information models (Mankiw and Reis, 2002).
15
contrast are quite complicated with dynamic properties emerging from theinteraction of pricing with many other features of the model. The model wepresent is in its log-linearised version (see Ascari 2003, Dixon and Kara 2005for the derivation from microeconomic foundation).To model the demand side, we use the Quantity Theory11:
yt = mt � pt
where (pt; yt) are aggregate price and output and mt the money supply. Wemodel the monetary growth process as an autoregressive process of order oneAR (1) :
mt = mt�1 + "t
"t = �"t�1 + �t
where �t is a white noise error term (e¤ectively a monetary growth shock).Following CEE we set � = 0:5:The optimal �exible price p�t at period t in all sectors is given by:
p�t = pt + yt (21)
The key parameter captures the sensitivity of the �exible price to output12.As discussed in Dixon and Kara (2010), there are a range of calibrated andestimated values for : for illustrative purposes, we use the "moderate" caseof = 0:1 as in Mankiw and Reis (2002). As discussed in Ascari (2003) andEdge (2002), the value of can be interpreted as resulting from either wageor price-setting. We therefore report the results using both the French wageand price data.Knowing (21) we can use the GTE price-setting equations and price for-
mulae (2); (3); (4) :to derive actual price-setting. We can do the same for theGC price-setting equations (5); (6); (7). To calibrate the model parameters�iw and hi; we use the micro data estimates presented above in section 3. Inthe case of the Calvo model, we simply take the GC and have a constant haz-ard �h taken from the data. We now take this simple quantity theory (QT)
11In the case of � = 0 below, the quantity theory can be seen as resulting from an Eulerequation (see Ascari 2003).12This can be due to increasing marginal cost and/or an upward sloping supply curve
for labour. See for example Walsh (2003, chapter 5) and Woodford (2003, chapter 3).
16
framework and subject it to a pure one-o¤ monetary growth shock �t > 0,which dies away rapidly with � = 0:5. The cumulative e¤ect of the shockin the limit is twice the initial shock. The model, as well as that of nextsection, is solved and simulations are performed using the DYNARE toolbox(Juillard, 1996). In Figure 3, we depict the impulse response functions foroutput and in�ation.
Insert FIGURE 3
There are two main observations to be made. First, in the in�ation IRF,there is no hump shape in either the Calvo or the GCE model, but thereis a hump shape with the GTE. This result con�rms, in a set-up that usesdata on actual distributions of price durations, the �nding of Dixon andKara (2010). Second, both the GTE and the GCE predict a more persistentin�ation and output response than the simple Calvo model.The intuition behind the hump is that in the GTE, �rms that are re-
setting their price are less forward looking on average in their pricing decisionthan in Calvo. That is because they know exactly how long their spell willlast, and so can ignore what happens after the spell �nishes (since they willbe able to choose another price). For example, the �rms with one periodspells only look at what is happening in the current period. That meansthat they will raise their prices less than �rms who have longer spells andso are more forward looking and anticipate future in�ation that will occurduring the spell and hence raise their price by more in anticipation of this.In the GCE and Calvo framework, all �rms that reset their prices have tolook forward F periods, since there is a possibility that their price might lastthat long. This means that the Calvo and GCE �rms raise their prices moston impact.The GC and GTE are more persistent for both in�ation and output than
Calvo. The intuition here is that the French price data has a fatter tail of longspells in the distribution of durations (and the cross-sectional DAF) than ispresent in the Calvo distribution. As shown in Dixon and Kara (2005), thatthe presence of long-contracts has a disproportionate e¤ect on the behaviorof aggregate output and in�ation due to the strategic complementarity ofprices13.
13See also Carvalho (2006) in the context of sectoral heterogeneity using the Calvoapproach.
17
4.1.1 Wage rigidity in the QT model.
We can do the same exercise calibrating the Calvo, GC and GT models withthe wage data. We should note however that the wage data does not havea long fat tail: indeed after 4 quarters, the proportion of long-spells is lowerin the data than in the Calvo distribution. We would therefore expect to seethe Calvo model as no less persistent than the GTE or GC:In Figure 4, we depict the impulse responses for all three models using
the wage data.Insert FIGURE 4
As we see, the in�ation and output IRFs for the Calvo and GC are verysimilar (and indeed both very di¤erent from the GTE case). There is anin�ation hump for the GTE, with an impact e¤ect on wage in�ation beingless than in the other two cases: but from the second quarter onwards thee¤ect on wage in�ation is larger. This is mirrored in the output IRF: thereis initially a greater e¤ect on output under the GTE, but after the thirdquarter there is less.If we consider the simple QT framework, we can see that the nature of the
empirical distribution matters. We have taken two distributions from themicro-data for the same economy: that of wages and prices. Whilst there aresome qualitative similarities, the exact shape of the distribution matters. Inparticular if we take the GC and the GT , they may give rise to similar IRFsfor output (in the case of price-data) or not (wage-data). This suggests thatthe micro-evidence is needed to evaluate the respective merits of the models.
4.2 A DSGE model: Smets and Wouters (2003)
In this section, we use the Smets and Wouters (2003) model, a now standardmodel of the euro area widely used for monetary policy analysis. We writeit down in its log-linearized form, which is for convenience reported in theappendix. The SW model is much more complicated than the simple QTmodel we have just used. There are many sources of dynamics other thanprices and wages: capital adjustment (and capital utilization), consumerdynamics with habit formation, and a monetary policy reaction function.The behavior of the model is the outcome of the interaction of all of theseprocesses together as it should be in a DSGE model. Hence the e¤ectof pricing dynamics is not isolated as in the simple QT framework of theprevious section.
18
4.2.1 Embedding GT and GC set-up in Smets and Wouters
Our strategy is the following. We are going to alter the structure of bothprice and wage rigidity in the model. We �rst remove the price and wagein�ation NKPC 0s from the SW model: that is equations (32-33) of theoriginal article. The rest of the model is left as it is. We then replace thesewith the nominal price and wage equations we derived in section 2, and de�neprice in�ation as the di¤erence in prices �t = pt� pt�1 and wage in�ation as�wt = wt � wt�1:To describe the price-setting decision, we can de�ne (nominal) marginal
cost in terms of the rental on capital and nominal wages
mct = (1� �)wt + �rkt � "at (22)
where rkt is the rental rate of capital and "at a productivity shock. Hence, in
log-linear form we have the optimal �ex-price equation
p�t = mct (23)
We can then use (23) to directly implement theGTE price equations (1); (2); (3; )and also the wage equations (11); (12); (13); (15); (16); (17).Similarly, we can use (23) to implement the GC price equations (6); (7)
and wage equations (19); (20) : To implement the Calvo model, we simplytake the GC model and set the reset-probability constant and equal to �h forprices and �hw for wages14.We underline that following our approach of starting from the micro-data
evidence, we remove indexation (which is a strong mechanism for creatingpersistence) from the SW model. We can then see how the price and wageequations without indexation but re�ecting the micro-data perform. We donot seek to re-estimate the SW model in this paper: our purpose is not toestimate a DSGE model of the Euro area. Rather, we want to illustrate howeasy it is to introduce evidence from the micro-data into a complex DSGEmodel such as the widely used SW model. Hence we take the calibratedor estimated values for parameters directly from the SW paper. For thoseparameters that were estimated in SW; we retain the mode of the posteriordistribution for each parameter (values are listed in the appendix).
14There is some approximation here, as we are truncating the Calvo distribution. How-ever, the di¤erence is quantitatively negligible: we ran the original code for the SW model(with the NKPC in terms of price and wage in�ation) with zero-indexation and found novisible di¤erence.
19
4.2.2 Monetary policy shock under GT and GC price and wagecontracts.
Figure 5 reports the IRF for in�ation and output in the SW model with GTand GC contracts following a monetary shock. We see that in this far morecomplex model, we get pretty much the same conclusions as in the simpleQT model. First, in�ation and output are more persistent for the GTE andGCE than with the Calvo set-up. Second, there is a hump-shaped responseof in�ation for the GTE, whilst the GC and Calvo have initial peak impact.
Insert FIGURE 5
The timing of the in�ation peak is earlier than in the original SW model:with the GTE it is 3 quarters, whilst in SW it is 5 quarters. It is however notsurprising that the model is not able to reach the same degree of persistenceas the original model. First, we are not re-estimating the model, and usea set of auxiliary parameters that were estimated to �t the data under theCalvo-with-indexation assumption. Re-estimating the full model, with theGTE or GCE assumption on euro area data would probably come closerto �tting the actual response of in�ation to monetary policy shock. Second,we have removed the indexation assumption both for wage and prices: Oneof the main roles of indexation is to generate a hump shaped response ofin�ation. Overall, the fact that we get a hump with the GTE even in thecomplicated SW framework shows that this is a robust result. Conversely,the fact that the GC does not give us a hump is also shown to be robust.
4.2.3 Technology shock
We also consider the case of a productivity shock and corresponding IRFin Figure 6. The shock is a persistent but non-permanent increase in totalfactor productivity. After the shock, there is an initial decline in marginalcost leading to a fall in prices and negative in�ation for the �rst 5 quarters.This is followed by positive in�ation as the shock dies away. Contrastingwith a quantity theory model, but in accordance with the standard Smetsand Wouters model, the long run impact on prices and wages is non-zero:the speci�c monetary policy rule employed results in a fall in the level ofprices and wages, of about a third in absolute value of the maximum short-run e¤ect. The e¤ect on output is everywhere positive, peaking at 7 quartersand very gradually dying away.
20
The di¤erences between the alternative price-setting models depend onhow they balance prices/in�ation and output over this path. As in the caseof a monetary shock, the impact e¤ect on prices is smaller for the GTE thanthe models where �rms/unions do not know the length of the price spell (GCand Calvo). However, all three models are quite similar in terms of the shapeand position of the IRF, unlike the case of the monetary shock. This is due tothe fact that the trajectory of the general price level is non-monotonic. In theGTE economy, the same mecanism as for the monetary policy shock plays arole in explaining a dampened reaction of the price level. In the case of theCalvo and GC economies, all price-setters have to consider the likelihood ofa long-price-spell. At a longer horizon however, due to the price level tendsto go back to its initial level, the required increase in price is smaller. As aresult, the impact e¤ect for both type of models is relatively close.
Insert FIGURE 6
5 Conclusion
In this paper, we have shown how we can take the micro-data on pricesand wages seriously and introduce them directly into our analysis of macro-economic policy using the standard DGSE models used today. Using thetheoretical framework of Dixon (2009), we have shown how we can take theestimated hazard function as a representation of the distribution of price-spell durations in the data and use it to infer the cross-sectional distributionunder the assumption of a steady-state. From these ways of looking at themicro-data, we can think of price and wage-setting models that are directlyconsistent with the micro-data: the Generalized Calvo and Generalized Tay-lor models of pricing. Also, for the �rst time to our knowledge, we showhow we can do this not only for prices or wages on their own but for bothwages and prices. We are able to use French original micro data to calibrateseparately wage and price setting and combine them in a consistent DGSEapproach.Perhaps the most interesting result we �nd is that if we adopt the Gener-
alized Taylor approach in both the output and labour market, we are able togenerate a hump-shaped response of in�ation to a monetary shock. This isnot so in the case of the generalized Calvo approach. This generalizes Dixonand Kara (2010) for an actual distribution of wage and price durations from
21
the euro area in a realistic model. In the case of a productivity shock, we�nd that all three approaches lead to a quite similar response.There are of course many ways to move on from this exercise. First, we
might choose to re-estimate the SW model with the wage and price-settingmodels derived from the micro-data. The micro-data used here could provideeither calibrated parameters of the pricing block or an initial distribution foreuro area parameters in the context of a Bayesian estimation. However, sincethe SW and CEE models were developed with di¤erent pricing models, itmight well be that we would want to change the structure of the models insome ways in addition to the pricing part. Second, we could undertake anoptimal policy exercise within this framework. Kara (2010) has conducteda comparison of optimal policy with a GTE in the simple quantity theorysetting: he �nds that the optimal policy with a GTE is similar to that derivedunder Calvo pricing. It would be interesting to see how this carries over tothe more complicated SW approach in this paper. These remain for futurework.
6 References
Alvarez LJ, Burriel P, Hernando I (2005). Do decreasing hazard functionsof price durations make any sense? ECB working paper series, No.461.
Ascari G. (2000), Optimizing Agents, Staggered wages and the persistencein the real e¤ects of monetary shocks, Economic Journal, 110, 664-686.
Ascari G. (2003), Price and Wage Staggering: a Unifying framework, Jour-nal of Economic Surveys, 17, 511-540.
Ascari G (2004), Staggered prices and trend in�ation: some nuisances, Re-view of Economic Dynamics, 7, 642-667.
Avouyi-Dovi S., Fougère D, Gautier E (2010)Wage Rigidity, Collective Bar-gaining and the Minimum Wage: Evidence from French agreementsdata, mimeo,Banque de France
Baudry L, Le Bihan H, Sevestre P and Tarrieu S (2007). What do thirteenmillion price records have to say about consumer price rigidity? OxfordBulletin of Economic Statistics, 69, 139-183.
22
Blanchard O. and Kiyotaki N. (1987), "Monopolistic Competition and theE¤ects of Aggregate Demand," American Economic Review 77, 647-66
Carvalho, C. (1995) "Firmas Heterogeneas, Sobreposicao de Contratos eDesin�acao", Pesquisa e Planejamento Economico, V. 25 n. 3, p. 479-496.
Carvalho, C (2006) "Heterogeneity in Price Stickiness and the Real E¤ectsof Monetary Shocks," Frontiers of Macroeconomics: Vol. 2 : Iss. 1,Article 1.
Christiano L., Eichenbaum M. et Evans C, (2005), Nominal Rigidities andthe Dynamic E¤ects of a Shock to Monetary Policy, Journal of PoliticalEconomy, 113, 1-45
Coenen G, Levin AT, Christo¤el K (2007), Identifying the in�uences ofnominal and real rigidities in aggregate price-setting behavior, Journalof Monetary Economics, 54, 2439-2466
Cogley T. and Sbordone A. (2008) Trend In�ation, Indexation, and In�ationPersistence in the New Keynesian Phillips Curve. American EconomicReview, 98(5).
Dhyne, E., Alvarez, L., Le Bihan, H., Veronese, G., Dias, D., Ho¤man, J.,Jonker, Lünneman, P., Rumler, F. and Vilmunen, J. (2006), �PriceChanges in the Euro Area and the United States: Some Facts fromIndividual Consumer price Data,� Journal of Economic Perspectives,20, 171-192.
Dixon H. (2009): "A uni�ed framework for understanding and comparingdynamic wage and price setting models", Banque de France WorkingPaper 259.
Dixon, H. and Kara, E (2005): "Persistence and nominal inertia in a gen-eralized Taylor economy: how longer contracts dominate shorter con-tracts", European Central Bank Working Paper 489, forthcoming Eu-ropean Economic Review.
Dixon, H and Kara, E (2006): "How to Compare Taylor and Calvo Con-tracts: A Comment on Michael Kiley", Journal of Money, Credit andBanking., 38, 1119-1126..
23
Dixon H, Kara E (2010): Can we explain in�ation persistence in a way thatis consistent with the micro-evidence on nominal rigidity, Journal ofMoney, Credit and Banking, 42(1), 151-170.
Edge, R. 2002 "The Equivalence of Wage and Price Staggering in MonetaryBusiness Cycle Models." Review of Economic Dynamics, 5, 559�585.
Erceg C, Henderson D and Levin A (2000), Optimal Monetary Policy withStaggered Wage and Price Contracts, Journal of Monetary Economics,46, 281-313.
Fougère D, Le Bihan H and Sevestre P(2007) Heterogeneity in price stick-iness: a microeconometric investigation, Journal of Business and Eco-nomics Statistics, 25(3), 247-264.
Guerrieri, L. 2006. "The In�ation Persistence of Staggered Contracts,"Journal of Money, Credit and Banking, 38(2), 483-494.
Heckel T., Le Bihan H., Montornes J.(2008) Sticky wages: evidence fromquarterly microeconomic data ECB Working Paper No. 893.
Juillard M, (1996) Dynare : A Program for the Resolution and Simula-tion of Dynamic Models with Forward Variables Through the Use of aRelaxation Algorithm, CEPREMAP Working Paper 9602.
Kara E (2010), Optimal Monetary Policy in the Generalized Taylor Econ-omy (forthcoming, Journal of Economic Dynamics and Control).
Klenow, P., and Malin B., (2010). "Microeconomic Evidence on Price-Setting", forthcoming, Handbook of Monetary Economics, Elsevier.
Mankiw N.G. and Reis R (2002): Sticky information versus sticky prices: aproposal to replace the new Keynesian Phillips curve, Quarterly Jour-nal of Economics, 117(4), 1295-1328.
Smets F and Wouters R, (2003). An Estimated Dynamic Stochastic Gen-eral Equilibrium Model of the Euro Area, Journal of the EuropeanEconomic Association,1, 1123-1175, 09
Taylor, J B (1993), Macroeconomic Policy in a World Economy, Norton.
Walsh C (2003) Monetary Theory and Policy (second edition), MIT Press.
24
Wolman A (1999). Sticky prices, marginal cost and the behavior of in�a-tion", Federal reserve bank of Richmond quarterly, 85, 29-47.
Woodford M (2003) Interest and Prices: Foundations of a Theory of Mon-etary Policy, Princeton.university press.
25
7 Appendix.
7.1 Deriving the reset wage in a GT economy.
Starting from (14), we �rst substitute for w�t+k using (13), and then substitutefor n(h)t+k using (8) and noting that w(h)t+k = xit for k = 0:::(i� 1) :
xit =1Pi�1k=0 �
k
i�1Xk=0
�kw�t+k
=1Pi�1k=0 �
k
i�1Xk=0
�kEt
�pt+k + �Ln(h)t+k +
�c1� b
(ct+k � b:ct+k�1)
�
=1Pi�1k=0 �
k
i�1Xk=0
�kEt
�pt+k + �L (�w (wt+k � xit) + nt+k) +
�c1� b
(ct+k � b:ct+k�1)
�Hence we can express the optimal reset wage in sector i as a function of
the aggregate variables fpt+k; wt+k; nt+k; ct+k; ct+k�1g only:
xit =1
(1 + �L�w)Pi�1
k=0 �k
i�1Xk=0
�kEt
�pt+k + �L (�wwt+k + nt+k) +
�c1� b
(ct+k � b:ct+k�1)
�
7.2 The log-linearized Smets-Wouters model and pa-rameter values.
First, there is the consumption Euler equation with habit persistence:
ct =b
1� bct�1 +
1
1 + bct+1 �
1� b
(1 + b)�c(rt � Et�t+1) +
1� b
(1 + b)�c"bt
Second there is the investment equation and related Tobin�s q equation
bIt =1
1 + �bIt�1 + �
1 + �EtbIt+1 + '
1 + �qt + "It
qt = � (rt � Et�t+1) +1� �
1� � + �rkEtqt+1 +
�rk
1� � + �rkEtr
kt+1 + �Qt
where , bIt is investment in log-deviation, qt is the shadow real price of capital,� is the rate of depreciation, �rk is the rental rate of capital. In addition, ' is
26
a parameter related to the cost of changing the pace of investment, and� ful�lls � =
�1� � + �rk
��1.
Capital accumulation is given by
bKt = (1� �) bKt�1 + � bIt�1Labour demand is given by
nt � bLt = � bwt + (1 + )brKt + bKt�1
Good market equilibrium condition is given by
bYt = (1� �ky � gy)bct + �kybIt + gyb"gt= �b"at + �� bKt�1 + �� brKt + �(1� �)bLt
The monetary policy reaction function is:
bit = �bit�1 + (1� �)f�t + r�(b�t�1 � �t) + rY (bYt � bY Pt )g
+f(r��(b�t � b�t�1) + r�Y ((bYt � bY Pt )� (bYt�1 � bY P
t�1))g+ �Rt
Shocks follow autoregressive processes:
"at = �a"at�1 + �at
"bt = �b"bt�1 + �bt
"It = �I"It�1 + �It
"Qt = �Q"Qt�1 + �Qt
"gt = �g"gt�1 + �gt
Note in the paper we focus on the e¤ects of two shocks: the monetarypolicy shock �Rt and the technology shock "
at : The calibration of the parame-
ters is given in Table A.1. below. It is based on the mode of the posteriorestimates, as reported in Smets and Wouters (2003).
27
Table A.1Parameter Value Interpretation� 0.99 Discount rate� 0.025 Depreciation rate� 0.30 Capital share�w 0.5 Mark-up wage'�1 6.771 Inv. adj. cost�c 1.353 Consumption utility elasticityb 0.573 Habit formation�L 2.400 Labor utility elasticity� 1.408 Fixed cost in production�e 0.599 Calvo employment 0.169 Capital util. adj. cost
Reaction function coe¢ cientsr� 1.684 to in�ationr�� 0.140 to change in in�ation� 0.961 to lagged interest ratery 0.099 to the output gapr�y 0.159 to change in the output gap
�a 0.823 persistence, productivity shock
28
02
46
810
1214
160
0.050.
1
0.150.
2
0.250.
3
0.350.
4
Haz
ard
func
tion
Pric
es (
Mon
thly
freq
uenc
y)
Fig
ure
1
02
46
810
1214
160.
250.3
0.350.
4
0.450.
5
0.550.
6
0.650.
7
0.75
Haz
ard
func
tion
Wag
es (
Qua
rter
ly fr
eque
ncy)
02
46
810
1214
160
0.050.
1
0.150.
2
0.250.
3
0.350.
4
Dis
trib
utio
n of
dur
atio
ns a
nd D
AF
Pric
es (
Mon
thly
freq
uenc
y)
Fig
ure
2
02
46
810
1214
160
0.1
0.2
0.3
0.4
0.5
0.6
Dis
trib
utio
n of
dur
atio
ns a
nd D
AF
Wag
es (
Qua
rter
ly fr
eque
ncy)
02
46
810
1214
160.
2
0.4
0.6
0.81
1.2
1.4
1.6
Out
put
Cal
voG
CE
GT
E
Fig
ure
3
02
46
810
1214
160
0.050.
1
0.150.
2
0.25
Infla
tion
IRF
of m
oney
gro
wth
sho
ck in
Qua
ntity
The
ory
mod
el fo
r G
TE
, GC
E, C
alvo
pric
e−se
tting
02
46
810
1214
160
0.2
0.4
0.6
0.81
Out
put
Cal
voG
CE
GT
E
Fig
ure
4
02
46
810
1214
160
0.050.
1
0.150.
2
0.250.
3
0.35
Infla
tion
IRF
of m
oney
gro
wth
sho
ck in
Qua
ntity
The
ory
mod
el fo
r G
TE
, GC
E, C
alvo
sch
eme
− W
age
calib
ratio
n
02
46
810
1214
16−
0.35
−0.
3
−0.
25
−0.
2
−0.
15
−0.
1
−0.
050O
utpu
t
Cal
voG
CE
GT
E
Fig
ure
5
02
46
810
1214
16−
0.25
−0.
2
−0.
15
−0.
1
−0.
050In
flatio
n
IRF
for
mon
etar
y po
licy
shoc
k in
the
Smet
s an
d W
oute
rs m
odel
with
GT
E, G
CE
and
Cal
vo p
rice
/wag
e se
tting
02
46
810
1214
16−
0.4
−0.
20
0.2
0.4
0.6
0.81
Out
put
Cal
voG
CE
GT
E
Fig
ure
6
02
46
810
1214
16−
1
−0.
8
−0.
6
−0.
4
−0.
20In
flatio
n
IRF
for
tech
nolo
gy s
hock
in S
met
s an
d W
oute
rs m
odel
GT
E, G
CE
, Cal
vo p
rice
/wag
e se
tting
CESifo Working Paper Series for full list see Twww.cesifo-group.org/wp T (address: Poschingerstr. 5, 81679 Munich, Germany, [email protected])
___________________________________________________________________________ 3057 Joseph Plasmans and Ruslan Lukach, The Patterns of Inter-firm and Inter-industry
Knowledge Flows in the Netherlands, May 2010 3058 Jenny E. Ligthart and Sebastian E. V. Werner, Has the Euro Affected the Choice of
Invoicing Currency?, May 2010 3059 Håkan Selin, Marginal Tax Rates and Tax-Favoured Pension Savings of the Self-
Employed – Evidence from Sweden, May 2010 3060 Richard Cornes, Roger Hartley and Yuji Tamura, A New Approach to Solving
Production-Appropriation Games with Many Heterogeneous Players, May 2010 3061 Ronald MacDonald and Flávio Vieira, A Panel Data Investigation of Real Exchange
Rate Misalignment and Growth, May 2010 3062 Thomas Eichner and Rüdiger Pethig, Efficient Management of Insecure Fossil Fuel
Imports through Taxing(!) Domestic Green Energy?, May 2010 3063 Vít Bubák, Evžen Kočenda and Filip Žikeš, Volatility Transmission in Emerging
European Foreign Exchange Markets, May 2010 3064 Leonid V. Azarnert, Après nous le Déluge: Fertility and the Intensity of Struggle against
Immigration, May 2010 3065 William E. Becker, William H. Greene and John J. Siegfried, Do Undergraduate Majors
or Ph.D. Students Affect Faculty Size?, May 2010 3066 Johannes Becker, Strategic Trade Policy through the Tax System, May 2010 3067 Omer Biran and Françoise Forges, Core-stable Rings in Auctions with Independent
Private Values, May 2010 3068 Torben M. Andersen, Why do Scandinavians Work?, May 2010 3069 Andrey Launov and Klaus Wälde, Estimating Incentive and Welfare Effects of Non-
Stationary Unemployment Benefits, May 2010 3070 Simon Gächter, Benedikt Herrmann and Christian Thöni, Culture and Cooperation, June
2010 3071 Mehmet Bac and Eren Inci, The Old-Boy Network and the Quality of Entrepreneurs,
June 2010 3072 Krisztina Molnár and Sergio Santoro, Optimal Monetary Policy when Agents are
Learning, June 2010
3073 Marcel Boyer and Donatella Porrini, Optimal Liability Sharing and Court Errors: An
Exploratory Analysis, June 2010 3074 Guglielmo Maria Caporale, Roman Matousek and Chris Stewart, EU Banks Rating
Assignments: Is there Heterogeneity between New and Old Member Countries? June 2010
3075 Assaf Razin and Efraim Sadka, Fiscal and Migration Competition, June 2010 3076 Shafik Hebous, Martin Ruf and Alfons Weichenrieder, The Effects of Taxation on the
Location Decision of Multinational Firms: M&A vs. Greenfield Investments, June 2010 3077 Alessandro Cigno, How to Deal with Covert Child Labour, and Give Children an
Effective Education, in a Poor Developing Country: An Optimal Taxation Problem with Moral Hazard, June 2010
3078 Bruno S. Frey and Lasse Steiner, World Heritage List: Does it Make Sense?, June 2010 3079 Henning Bohn, The Economic Consequences of Rising U.S. Government Debt:
Privileges at Risk, June 2010 3080 Rebeca Jiménez-Rodriguez, Amalia Morales-Zumaquero and Balázs Égert, The
VARying Effect of Foreign Shocks in Central and Eastern Europe, June 2010 3081 Stephane Dees, M. Hashem Pesaran, L. Vanessa Smith and Ron P. Smith, Supply,
Demand and Monetary Policy Shocks in a Multi-Country New Keynesian Model, June 2010
3082 Sara Amoroso, Peter Kort, Bertrand Melenberg, Joseph Plasmans and Mark
Vancauteren, Firm Level Productivity under Imperfect Competition in Output and Labor Markets, June 2010
3083 Thomas Eichner and Rüdiger Pethig, International Carbon Emissions Trading and
Strategic Incentives to Subsidize Green Energy, June 2010 3084 Henri Fraisse, Labour Disputes and the Game of Legal Representation, June 2010 3085 Andrzej Baniak and Peter Grajzl, Interjurisdictional Linkages and the Scope for
Interventionist Legal Harmonization, June 2010 3086 Oliver Falck and Ludger Woessmann, School Competition and Students’
Entrepreneurial Intentions: International Evidence Using Historical Catholic Roots of Private Schooling, June 2010
3087 Bernd Hayo and Stefan Voigt, Determinants of Constitutional Change: Why do
Countries Change their Form of Government?, June 2010 3088 Momi Dahan and Michel Strawczynski, Fiscal Rules and Composition Bias in OECD
Countries, June 2010
3089 Marcel Fratzscher and Julien Reynaud, IMF Surveillance and Financial Markets – A
Political Economy Analysis, June 2010 3090 Michel Beine, Elisabetta Lodigiani and Robert Vermeulen, Remittances and Financial
Openness, June 2010 3091 Sebastian Kube and Christian Traxler, The Interaction of Legal and Social Norm
Enforcement, June 2010 3092 Volker Grossmann, Thomas M. Steger and Timo Trimborn, Quantifying Optimal
Growth Policy, June 2010 3093 Huw David Dixon, A Unified Framework for Using Micro-Data to Compare Dynamic
Wage and Price Setting Models, June 2010 3094 Helmuth Cremer, Firouz Gahvari and Pierre Pestieau, Accidental Bequests: A Curse for
the Rich and a Boon for the Poor, June 2010 3095 Frank Lichtenberg, The Contribution of Pharmaceutical Innovation to Longevity
Growth in Germany and France, June 2010 3096 Simon P. Anderson, Øystein Foros and Hans Jarle Kind, Hotelling Competition with
Multi-Purchasing: Time Magazine, Newsweek, or both?, June 2010 3097 Assar Lindbeck and Mats Persson, A Continuous Theory of Income Insurance, June
2010 3098 Thomas Moutos and Christos Tsitsikas, Whither Public Interest: The Case of Greece’s
Public Finance, June 2010 3099 Thomas Eichner and Thorsten Upmann, Labor Markets and Capital Tax Competition,
June 2010 3100 Massimo Bordignon and Santino Piazza, Who do you Blame in Local Finance? An
Analysis of Municipal Financing in Italy, June 2010 3101 Kyriakos C. Neanidis, Financial Dollarization and European Union Membership, June
2010 3102 Maela Giofré, Investor Protection and Foreign Stakeholders, June 2010 3103 Andrea F. Presbitero and Alberto Zazzaro, Competition and Relationship Lending:
Friends or Foes?, June 2010 3104 Dan Anderberg and Yu Zhu, The Effect of Education on Martial Status and Partner
Characteristics: Evidence from the UK, June 2010 3105 Hendrik Jürges, Eberhard Kruk and Steffen Reinhold, The Effect of Compulsory
Schooling on Health – Evidence from Biomarkers, June 2010
3106 Alessandro Gambini and Alberto Zazzaro, Long-Lasting Bank Relationships and
Growth of Firms, June 2010 3107 Jenny E. Ligthart and Gerard C. van der Meijden, Coordinated Tax-Tariff Reforms,
Informality, and Welfare Distribution, June 2010 3108 Vilen Lipatov and Alfons Weichenrieder, Optimal Income Taxation with Tax
Competition, June 2010 3109 Malte Mosel, Competition, Imitation, and R&D Productivity in a Growth Model with
Sector-Specific Patent Protection, June 2010 3110 Balázs Égert, Catching-up and Inflation in Europe: Balassa-Samuelson, Engel’s Law
and other Culprits, June 2010 3111 Johannes Metzler and Ludger Woessmann, The Impact of Teacher Subject Knowledge
on Student Achievement: Evidence from Within-Teacher Within-Student Variation, June 2010
3112 Leif Danziger, Uniform and Nonuniform Staggering of Wage Contracts, July 2010 3113 Wolfgang Buchholz and Wolfgang Peters, Equity as a Prerequisite for Stable
Cooperation in a Public-Good Economy – The Core Revisited, July 2010 3114 Panu Poutvaara and Olli Ropponen, School Shootings and Student Performance, July
2010 3115 John Beirne, Guglielmo Maria Caporale and Nicola Spagnolo, Liquidity Risk, Credit
Risk and the Overnight Interest Rate Spread: A Stochastic Volatility Modelling Approach, July 2010
3116 M. Hashem Pesaran, Predictability of Asset Returns and the Efficient Market
Hypothesis, July 2010 3117 Dorothee Crayen, Christa Hainz and Christiane Ströh de Martínez, Remittances,
Banking Status and the Usage of Insurance Schemes, July 2010 3118 Eric O’N. Fisher, Heckscher-Ohlin Theory when Countries have Different
Technologies, July 2010 3119 Huw Dixon and Hervé Le Bihan, Generalized Taylor and Generalized Calvo Price and
Wage-Setting: Micro Evidence with Macro Implications, July 2010
